# Odd behavior when logstash finished pipeline
**URL:** https://discuss.elastic.co/t/odd-behavior-when-logstash-finished-pipeline/114612
**Category:** Logstash
**Created:** [January 8, 2018, 8:34pm UTC](https://discuss.elastic.co/t/odd-behavior-when-logstash-finished-pipeline/114612 "2018-01-08T20:34:53Z")
**Posts on this page:** 5
**Page:** 1
### Author:  [@arisbanach](https://discuss.elastic.co/u/arisbanach)
#### Post date: [January 8, 2018, 8:34pm UTC](https://discuss.elastic.co/t/odd-behavior-when-logstash-finished-pipeline/114612/1 "2018-01-08T20:34:53Z")
I have two pipelines defined in `pipelines.yml`: one for my main pipeline which imports data from jdbc input, and another for the dead\_letter\_queue input.
When I run Logstash when it gets to the end of the database, it seems to stop on the last database query and doesn't import anything else. I thought Logstash was supposed to end the pipeline and restart after it was done or something. Is that not the normal behavior?
---
### Author:  [@guyboertje](https://discuss.elastic.co/u/guyboertje)
#### Post date: [January 11, 2018, 12:00pm UTC](https://discuss.elastic.co/t/odd-behavior-when-logstash-finished-pipeline/114612/2 "2018-01-11T12:00:42Z")
Logstash wants to run forever processing streams of events.
Some inputs set up a server like process that waits (forever) for clients to connect.
Some inputs read from local or networked (S3) folders in a loop looking for new files to read or more content in found files (tail).
Some inputs query message brokers in a loop fetching data from topics.
Some inputs use cron like schedules to periodically read from a source.
Only one input `generator` will run for a certain time and then shutdown Logstash.
Due to continual streaming, the concept of reaching the end of a pipeline only applies when the `generator` input is used.
---
### Author:  [@arisbanach](https://discuss.elastic.co/u/arisbanach)
#### Post date: [January 11, 2018, 2:10pm UTC](https://discuss.elastic.co/t/odd-behavior-when-logstash-finished-pipeline/114612/3 "2018-01-11T14:10:28Z")
So for JDBC input is it best practice to use the `schedule` parameter and not just let it run?
---
### Author:  [@guyboertje](https://discuss.elastic.co/u/guyboertje)
#### Post date: [January 11, 2018, 2:32pm UTC](https://discuss.elastic.co/t/odd-behavior-when-logstash-finished-pipeline/114612/4 "2018-01-11T14:32:04Z")
In most cases, yes.
However if you want to use the elasticsearch filter to do enhancement on another pipeline and you choose to load the records from a static reference DB table into ES then you would only run this "loader" pipeline once.
---
### Author:  [@system](https://discuss.elastic.co/u/system)
#### Post date: [February 8, 2018, 2:32pm UTC](https://discuss.elastic.co/t/odd-behavior-when-logstash-finished-pipeline/114612/5 "2018-02-08T14:32:10Z")
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.